VFX Graph won’t render Set Position from Map if VFX Window is not open in Editor

Hi @Elliott-Mitchell ,

Can you share a screenshot of your graph or a gif/video demonstrating the issue?
(or otherwise log an issue and we can take a look)

Cheers!

 

Hi @Elliott-Mitchell ,
Thanks for filing a bug, we'll take a look.

From your video, I see you have ticked all of the preferences. Many of them are for debug purposes, do you have a strong reason for needing them all on? The experimental blocks and the additional debug info are helpful for every day use, but if you don't have as reason for having the last 3 on, I'd recommend turning them off (which seems to also solve your issue).


But either way, the team will take a look at the issue you filed, thanks!

 

Hey @Elliott-Mitchell ,

I took a look and it indeed looks like an iOS bug.
We have a task on the public roadmap for adding mobile support for compute-capable devices, so thanks for logging this, we were not aware of this issue.

Until a fix is made, as a workaround I observed that the issue only happens in Initialize, so you can potentially move the Position from Map block down to Update or the Output.


If you move it to Update, you should change the Sample Mode to Random Constant, so particles get the same position each frame. This won't work with your gravity, however.

If you move it to Output, just change your block's composition from Overwrite to Add and remove the sphere position in Initialize, and it should work with the gravity as well.

This is what the min. graph change will look like for the Output workaround:


You can by the way emulate closing the VFX window by forcing the graph to use runtime shaders even if the VFX window is open (by default if the graph is open it uses an "edit mode" set which allows you to modify values and see the changes. The runtime mode is the optimized one that's doesn't allow that).


Hope this helps!

 

It looks like the particles are teleporting. There might be other blocks manipulating the position or velocity which are interfering, but it's hard to tell without looking at the graph. If you'd like, feel free to DM me the graph (or a min repro) and I can take a look. Otherwise I'm afraid I can't reproduce what you are seeing with the graph from the bug case.

 

Hello @Elliott-Mitchell, while looking into this issue, I noticed that the issue does not happen when adding a Value Bias different from 0. You can use a very small value here as a workaround, waiting for the issue to be fixed.
I hope it is not too late!